CKD Guardian is a kidney-focused digital health project built for CKD detection, monitoring, alerts, and telemedicine support.
It is designed as a mobile-first system with web compatibility, where patients can submit CKD readings, receive risk analysis, view alerts, get notifications, and interact with doctors through consultation workflows.
The project is based on a manually prepared and tested CKD dataset. Its ML logic works on the same structured clinical and sensor-based features used during training.
The objective of CKD Guardian is to provide a simple and practical platform for:
- CKD risk detection using structured medical inputs
- Monitoring kidney-related trends over time
- Generating alerts for abnormal readings
- Notifying doctors about urgent cases
- Supporting telemedicine consultations between doctors and patients
- Register and login
- Create patient profile
- Enter CKD readings manually
- Upload report image
- Extract report values into dataset-based fields
- View CKD risk score
- View latest readings and prediction trend
- Receive alerts
- Receive notifications
- View teleconsultation details
- View recent and past consultation history
- Doctor login
- Review urgent CKD cases
- View patient overview
- View recent and past patient overview records
- Receive urgent alerts and notifications
- Schedule consultations
- Add doctor advice
- Add prescription note
- Add patient instructions
- View recent and past consultation history
The app uses only the features included in the project dataset, such as:
- Creatinine
- Urine albumin
- ACR
- eGFR
- Systolic BP
- Diastolic BP
- Glucose
- Sensor values
- Symptom indicators
- Adherence score
Uploaded report images are used only to extract values and help autofill these structured fields.
The image itself is not directly used as a prediction input.
CKD app/
├── backend/
│ ├── app/
│ ├── requirements.txt
│ └── README.md
├── mobile-app/
│ ├── app/
│ ├── assets/
│ ├── package.json
│ ├── app.json
│ ├── eas.json
│ └── README.md
└── README.md
- React Native
- Expo
- React Navigation
- Axios
- Expo Web
- FastAPI
- SQLAlchemy
- Pydantic
- SQLite for local development
- PostgreSQL for deployment
- Joblib
- NumPy
- Scikit-learn
- Trained CKD model or heuristic fallback logic
- Brevo email integration
- In-app notification system
- Render for backend deployment
- Vercel for web deployment
- Expo EAS for Android/iOS builds
cd backend
source .venv/bin/activate
uvicorn app.main:app --reloadBackend API:
http://127.0.0.1:8000
Swagger docs:
http://127.0.0.1:8000/docs
cd mobile-app
npm install
npx expo start --localhostcd mobile-app
npx expo start --web- Backend deployed on Render
- Web frontend deployed on Vercel
- Android and iOS builds generated using Expo EAS
- PostgreSQL used as production database on Render
CKD Guardian uses a soft-clear workflow.
Recent patient overview → Clear → Past patient overview
Doctor overview data is not deleted permanently. Cleared overview records move to Past and remain visible.
Recent consultations → Clear → Past consultations
Consultation history is not deleted permanently. Cleared consultations move to Past and remain visible for both patient and doctor accounts.
- This project is focused on CKD dataset-driven prediction.
- It should not be presented as a general-purpose hospital AI system.
- Uploaded report images are used for value extraction and autofill only.
- The ML model works on structured dataset fields, not raw images.
- The project is intended for academic/project demonstration and workflow validation.
CKD Guardian Project Team